Vol. 6, 1920 PHYSICS: DUANE AND PATTERSON 
513 
of the wave-numbers, 1/X, orof the ratios of the frequencies of vibration to 
the Rydberg fundamental frequency for heavy atoms. In his classical 
memoirs on the wave-lengths of characteristic X-rays Moseley called at- 
tention to the fact that the square root of the wave-number (or frequency) 
of a line in the X-ray spectrum of an element is very nearly a linear func- 
tion of its atomic number. To test this relation in the case of critical 
TABLE 2 
Critical Absorption Wave-Numbers. L Series of X-Rays 
Values of i /X and of V 1 /X 
CHEMICAL 
ElvEMENT 
ATOMIC 
NUMBER 
i/X 
ViTx 
Lai 
DIFFERENCES 
J. Uli^bLCli 
7zl 
0 
8240 ±1 
0 
9077 
Platinum 
78 
0 
9341 =t 3 
0 
9665 
0 
0147=1= 1 
LtOIQ 
TO 
/y 
0. 9631^3 
0 
9814 
0 
0149 =^3 
Mercury 
oU 
0 
9933 ±4 
0 
9966 
0 
0152=1=4 
i naiium 
1 
0229 ±3 
1 
0114 
0 
0148±4 
IvCad 
CO 
1 
0530 ±3 
1 
0262 
0 
0148=^3 
Bismuth 
83 
1 
0851^3 
1 
0417 
0 
0155 =t=3 
Thorium 
90 
1 
3164=fc5 
1 
1 
i't / 0 
A 
\J 
ni f^i ± 1 
Uranium 
92 
1 
3862 ±5 
1 
1774 
0 
0151 =t2 
Tungsten 
74 
0 
9323 ±5 
0 
9656 
Platinum 
78 
1 
0728^4 
1 
0358 
0 
0176=^1 
Gold 
79 
1 
1120 =i=4 
1 
0545 
0 
0187 =t 4 
Mercury 
90 
1 
1494 ±4 
1 
0721 
0 
0176 ±4 
Thalium 
81 
1 
1883 ±5 
1 
0901 
0 
0180 ±5 
Lead 
82 
1 
2295 ±5 
1 
1088 
0 
0187=^5 
Bismuth 
83 
1 
2703 ±4 
1 
1271 
0 
0182:^5 
Thorium 
90 
1 
5908 ±8 
1 
2613 
0 
0192 ±1 
Uranium 
92 
1 
6898 ±8 
1 
2999 
0 
0193=^3 
Tungsten 
74 
0 
977 ±3 
0 
Las 
988 
Platinum 
78 
1 
1255^12 
1 
0609 
0 
0182 =b 5 
Gold 
79 
1 
162 ±1 
1 
0780 
0 
0171 =tll 
Mercury 
80 
1 
200 ±1 
1 
0953 
0 
0174=t=ll 
Thalium 
81 
1 
242 ±2 
1 
1142 
0 
0189=1=15 
Lead 
82 
1 
282 =t 2 
1 
1320 
0 
0178 =b 16 
Bismuth 
83 
1 
.328=^2 
1 
1523 
0 
0203 ±13 
Thorium 
90 
1 
.655^2 
1 
2863 
0 
0192=^2 
Uranium 
92 
1 
.759 ±2 
1 
3263 
0 
0200 =t= 7 
absorption data we have calculated the wave-numbers and the square 
roots of the wave-numbers for all of our values. Column 3, in Table 2, 
contains the wave-numbers and Column 4 their square roots. In Column 
5 we have tabulated the increase in the value of V i /X per unit increase in 
the atomic number. It appears that these differences do not vary from 
